基于卡尔曼滤波的短波等幅报降噪技术研究
|
摘要
本文研究的是短波莫尔斯电报信号的噪声抑制处理技术。在无线通信系统中,噪声信号的干扰是影响通信效果的主要原因。由于短波莫尔斯电报通信属于模拟通信的制式,依赖电离层天波反射和地波传输,对信道依赖性较强,随着工业电气化的发展,短波信道受到的干扰日益严重,莫尔斯信号传输质量急剧恶化,导致电台值班抄收水平日益下降,给从事人工守听抄收莫尔斯电报信号的报务员带来较大负担。本文研究的目的是使用数字信号处理技术,在不改动原有电台设备硬件结构的基础上,搭建在线的信号处理模块作为电台的后级音频处理系统,对电台输出的莫尔斯音频信号进行降噪处理后,提升莫尔斯音频信号的信噪比,改善莫尔斯音频信号的听感,在降低无线电报务值班人员劳动强度的基础上,提高值班电报信号抄收的准确率,提升电台值班业务水平。
在前人研究工作的基础上,本文首先对莫尔斯电码自身的点划及字符间的相关规律进行了分析,研究了影响莫尔斯电报信号的噪声类型,讨论了噪声对莫尔斯信号的污染方式,分析了短波通信设备架设方式对莫尔斯电报信号噪声的影响,研究了电报抄收人员在噪声干扰状态下实现人工抗噪声抄收的原理。根据莫尔斯信号电码的自相关特性,分析了人脑在记忆和识别莫尔斯电报信号中的分析处理过程,研究了在发报手法不均状态下,抄收人员抗噪声处理的方法。分析了莫尔斯信号现有的降噪处理技术,建立了短波莫尔斯电报信号的空间状态模型,应用卡尔曼滤波算法对其进行了推导,建立了卡尔曼滤波模型。探讨了卡尔曼滤波在短波莫尔斯电报信号音频滤波方法,在CCS环境下编写了卡尔曼滤波算法的DSP程序,结合现有值班电台的音频输出接口,搭建了以DSP5402和AIC23为核心的数字信号降噪处理模块,使用AIC23对电台输出的音频信号进行模数采样,在DSP5402中使用卡尔曼滤波程序对莫尔斯音频信号进行降噪处理,处理后音频信号通过AIC23耳机接口输出给报务员抄收。
为了验证卡尔曼滤波处理对莫尔斯音频信号的降噪处理性能,本文模拟了一段受高斯白噪声污染的莫尔斯音频信号,使用计算输出信噪比的方法,将卡尔曼滤波器与传统的窄带滤波器进行了对比试验,实验结果表明,卡尔曼滤波算法对系统参数和噪声的统计特性估计模型决定了卡尔曼滤波对莫尔斯信号的降噪处理性能,在莫尔斯信号的信噪比较大时,卡尔曼滤波器完全能够达到窄带滤波器的降噪效果。
In this paper, the short-wave morse telegraph signal noise reduction technology is discussed. In the wireless communication system, noise interference is the main effect of communication. Because short-wave morse telegraph communication belongs to the analog communication standard, it depends on the sky wave ionospheric reflection and ground wave transmission, as well as the channel. With the development of industrial electrification, the interference the short-wave channel suffers is growing, the quality of the transmission of morse signal has a sharp deterioration. It leds to declining levels of radio duty collecting, it makes it very difficult for the operator to listen to the morse telegraph telegraph signals. The purpose of this paper is to use the digital signal processing technology to build a online signal processing module to be after-level of the radio audio processing system, on the basis of no alteration of the original radio equipment hardware. After denoising of morse audio signal out from radio, it can raise the SNR of the morse audio signal, improve the listening sense of the morse audio signal, on the basis of reducing the labor intensity of radio wireless telegraphy service operator, improve the accuracy of collecting duty, and enhance the business level of professional skill.
Based on the previous studies, this paper first analyze the dot-and-dash of morse and the rules and the correlation between characters, it also study the effects of type of noise of morse signal, discuss the pollution mode of the morse signal, analyze the influence of the noise of short-wave communication which the erection methods of equipment makes, it studys the principle of artificial collecting of noise can achieve by collecting telegraph officers in noise state. According to signal autocorrelation of morse code, it analyzes the signal processing of human brain in memory and recognition of morse telegraphy. It studys ways to deal with anti-noise of collecting staff with a uneven way in transmitter mode. It analyzes the existing noise reduction technology in morse signal, establish a short space state model of morse telegraph. It uses kalman filter algorithm to conduct the derivation, and kalman filter model was established.it investigates the audio filter method of kallman filtering in short-wave morse telegraphy signal and writes the DSP code in CCS environment, with the combine of existing audio output interface in duty stations, it builds a digital signal denoiseprocessing module with the core of DSP5402 and AIC23, uses the AIC23 to do the AD sampling of audio signals out from the radio, and use the kalman filter algorithm to conduct the denoising of morse audio signal in DSP5402, after processing, the audio signal output to the radio operator for collecting through the AIC23 headphone.
In order to verify the performance of the noise reduction of kalman filter to the morse audio signal, we simulated an audio signal morse white a gaussian noise, use the method of calculation of the output SNR to take a contrast experiment of kalman filter and the traditional narrow-band filter. The results show that it is estimation model which kalman filter algorithm make for system parameters and noise statistics decides the performance of kalman filter can achieve in noise reduction of morse signal. When SNR of morse signal is larger, kalman filter is fully able to meet the noise reduction effect of narrow-band filter
在前人研究工作的基础上,本文首先对莫尔斯电码自身的点划及字符间的相关规律进行了分析,研究了影响莫尔斯电报信号的噪声类型,讨论了噪声对莫尔斯信号的污染方式,分析了短波通信设备架设方式对莫尔斯电报信号噪声的影响,研究了电报抄收人员在噪声干扰状态下实现人工抗噪声抄收的原理。根据莫尔斯信号电码的自相关特性,分析了人脑在记忆和识别莫尔斯电报信号中的分析处理过程,研究了在发报手法不均状态下,抄收人员抗噪声处理的方法。分析了莫尔斯信号现有的降噪处理技术,建立了短波莫尔斯电报信号的空间状态模型,应用卡尔曼滤波算法对其进行了推导,建立了卡尔曼滤波模型。探讨了卡尔曼滤波在短波莫尔斯电报信号音频滤波方法,在CCS环境下编写了卡尔曼滤波算法的DSP程序,结合现有值班电台的音频输出接口,搭建了以DSP5402和AIC23为核心的数字信号降噪处理模块,使用AIC23对电台输出的音频信号进行模数采样,在DSP5402中使用卡尔曼滤波程序对莫尔斯音频信号进行降噪处理,处理后音频信号通过AIC23耳机接口输出给报务员抄收。
为了验证卡尔曼滤波处理对莫尔斯音频信号的降噪处理性能,本文模拟了一段受高斯白噪声污染的莫尔斯音频信号,使用计算输出信噪比的方法,将卡尔曼滤波器与传统的窄带滤波器进行了对比试验,实验结果表明,卡尔曼滤波算法对系统参数和噪声的统计特性估计模型决定了卡尔曼滤波对莫尔斯信号的降噪处理性能,在莫尔斯信号的信噪比较大时,卡尔曼滤波器完全能够达到窄带滤波器的降噪效果。
In this paper, the short-wave morse telegraph signal noise reduction technology is discussed. In the wireless communication system, noise interference is the main effect of communication. Because short-wave morse telegraph communication belongs to the analog communication standard, it depends on the sky wave ionospheric reflection and ground wave transmission, as well as the channel. With the development of industrial electrification, the interference the short-wave channel suffers is growing, the quality of the transmission of morse signal has a sharp deterioration. It leds to declining levels of radio duty collecting, it makes it very difficult for the operator to listen to the morse telegraph telegraph signals. The purpose of this paper is to use the digital signal processing technology to build a online signal processing module to be after-level of the radio audio processing system, on the basis of no alteration of the original radio equipment hardware. After denoising of morse audio signal out from radio, it can raise the SNR of the morse audio signal, improve the listening sense of the morse audio signal, on the basis of reducing the labor intensity of radio wireless telegraphy service operator, improve the accuracy of collecting duty, and enhance the business level of professional skill.
Based on the previous studies, this paper first analyze the dot-and-dash of morse and the rules and the correlation between characters, it also study the effects of type of noise of morse signal, discuss the pollution mode of the morse signal, analyze the influence of the noise of short-wave communication which the erection methods of equipment makes, it studys the principle of artificial collecting of noise can achieve by collecting telegraph officers in noise state. According to signal autocorrelation of morse code, it analyzes the signal processing of human brain in memory and recognition of morse telegraphy. It studys ways to deal with anti-noise of collecting staff with a uneven way in transmitter mode. It analyzes the existing noise reduction technology in morse signal, establish a short space state model of morse telegraph. It uses kalman filter algorithm to conduct the derivation, and kalman filter model was established.it investigates the audio filter method of kallman filtering in short-wave morse telegraphy signal and writes the DSP code in CCS environment, with the combine of existing audio output interface in duty stations, it builds a digital signal denoiseprocessing module with the core of DSP5402 and AIC23, uses the AIC23 to do the AD sampling of audio signals out from the radio, and use the kalman filter algorithm to conduct the denoising of morse audio signal in DSP5402, after processing, the audio signal output to the radio operator for collecting through the AIC23 headphone.
In order to verify the performance of the noise reduction of kalman filter to the morse audio signal, we simulated an audio signal morse white a gaussian noise, use the method of calculation of the output SNR to take a contrast experiment of kalman filter and the traditional narrow-band filter. The results show that it is estimation model which kalman filter algorithm make for system parameters and noise statistics decides the performance of kalman filter can achieve in noise reduction of morse signal. When SNR of morse signal is larger, kalman filter is fully able to meet the noise reduction effect of narrow-band filter
引文
[1]何非常.军事通信[M].北京:国防工业出版社,2000.
[2]高晋占.探空莫尔斯码自动识别[J].电子测量与仪器学报,2006.(8): 5-9.
[3]樊昌元,文斌.应用于民航空管的TPU莫尔斯译码器[J].电子测量与仪器学报,2004,(9):45-48.
[4]马威,张敬修,王虎帮.Morse电码自动译码系统[J].网络与信息技术,2007, 26(6):51-55.
[5]林瑞华.用微处理机对莫尔斯电码进行机器译码[J].通信与计算技术,1984,(3):126-128.
[6]高勤,都新学.莫尔斯信号的新型滤波方法[C].中国电子学会第七届学术年会论文集.2001.
[7]邓自立,王欣,高媛.建模与估计[M].北京:科学出版社,2007.
[8]皇甫堪,陈建文,楼生强.现代数字信号处理[M].北京:电子工业出版社,2003.
[9]易克初.语音信号处理[M].北京:国防工业出版社,2000.
[10]于宏毅,张贴等.手工莫尔斯报的一种新型自动收报算法及其手法识别算法[J].电子对抗,1995,11(4):317-320.
[11]岳喜才,郑崇勋.基于离散Gabor谱的短波电报信号检测.数据采集及信号处理,1995,11(4):317-320.
[12]林佳仕.短波通信中信号检测算法研究.哈尔滨工程大学硕士论文,2003.
[13]戴逸松.微弱信号检测方法及仪器.北京:国防工业出版社,1994,1.
[14]朱岩.强噪声下基于小波变换的语音增强研究.哈尔滨工程大学硕士论文,2003.
[15]王国权,薛申芳,金声震,宁书年.卫星自主导航中卡尔曼滤波算法改进与计算机仿真.计算机仿真,2004,(7):17-21.
[16]冯鹏,邹世开.卡尔曼滤波器UD分解滤波算法的改进.遥测遥控,2005.1.
[17]庄琼琼.多传感器组合卡尔曼滤波技术研究.北京航空航天大学硕士学位论文,2002.
[18]刘理.基于信号相位匹配原理的单频信号检测.西北工业大学学报,2005,(2): 35-41.
[19]孙轶源.利用信号相位匹配原理的正弦信号参数估计.自然科学发展,2001,11.
[20]杨祥龙,汪乐宇.一种强噪声背景下弱信号检测的非线性方法.电子与信息学报,1999,14(1):234-235.
[21]林佳仕.短波通信中信号检测算法的研究.哈尔滨工程大学硕士学位论文,2004.1.
[22]何立刚.噪声背景下短波莫尔斯信号的自动检测和识别研究.哈尔滨工程大学硕士论文,2005.
[23]袁文澹,唐伟.微机模拟自动拍发MORSE电报[J].计算技术与自动化,2001,2.
[24]耿振华.用单片机制制作多功能莫尔斯码电路[J].电子报,2005,05-08.
[25] [美]阿姆巴德.信号、系统与信号处理[M].冯博琴等译.机械工业出版社,2001.
[26]程佩青.数字信号处理教程[M].北京:清华大学出版社,2001.
[27]李跃进.短波电台销噪技术的实现方法[J].通信与广播电视,2006,(2):124-127.
[28]霍建,刘鸿雁,段秀铭.手工拍发的莫尔斯码信号种类识别算法设计[J].鞍山科技大学学报,2006,29(4):351-355.
[29]林金朝,李国军,周晓娜,曾理.强噪声背景下高频CW电报信号的自动检测[J].山西电子技术,2008,10.39-40.
[30]李春晓,赵旦峰,李强.用语音识别技术实现莫尔斯报的自动识别[J].信息技术.2006,(2):34-39.
[31]陈莹莹,彭仕勇,高勇.莫尔斯报务信号的检测和识别[J].计算机应用与软件,2009,26(6):241-243.
[32]潘德惠.卡尔曼滤波浅述[J].信息与控制,1982,(1):42-46.
[33]张正林,谭文生.从无线通信的发展历程看战术短波通信的技术体制和发展趋势[J].陕西烽火通信,2002,(1):1-2.
[34]张莉,张骏凌,龚耀寰.软件无线电数字中频技术[J].电讯技术,1999,(6):48-51.
[35]杨路刚,陈斌,王永斌,喻莉.基于MSP430的莫尔斯报训练装置设计[J].现代电子技术,2009,(9):8-10.
[36]高晋占.探空莫尔斯码自动识别[J].电子测量与仪器学报,2006,20(4):5-9.
[37]冯锡生.短波通信电路设计.北京:国防工业出版社,1985.
[38]王敬东,申敏.TI TMS320C55x DSP在TD-SCDMA系统中的应用[J].重庆邮电学院学报,2002,14(2):22-25.
[39]张汝波,何立刚,李雪耀.强噪声背景下莫尔斯信号的自动检测与识别[J].哈尔滨工程大学学报,2006, 27 (1): 113-117.
[40]王金龙,沈良,任国春等.无线通信系统的DSP实现[M].北京:人民邮电出版社,2002.
[41]周学科,张太镒,王小军.基于软件无线电的短波中频数字化[J].重庆邮电学院学报,2004,16(4):34-37.
[42]李双田,李昌立,陈丹平,袁津润.短波数字接收机的算法研究和DSP实现[J].电子学报,1999,(7):27-34.
[43] R. M. Lober.A DSP-based approach to HF receiver design:Higher Performance at a Lower Cost. TECHNICAL SYMPOSIUM,1993
[44] R.J.Coy, C.N.Smith and P.R.Smith.Use of DSP within a high performance HF band receiver.In:Proc IEE Conference of Radio Receivers and Associated Systems, July,1990.
[45] Qian Shie, Chen Dapang. Optimal biorthogonal analy-sis window function for discrete Gabor transform. IEEE Trans Signal Processing, 1994,4.
[46] MITOLA J. The software radio architecture [J].IEEE Communication Magazine,1995, 33.
[47] Castella,F.R.,An adaptive two-dimensional Kalman tracking filter,IEEE Trans. Aerosp, Election Syst.,Vol,AES-16,
[48] Taxas Instruments Incorporated,TMS320 Assembly Language Programmer’s Guide, 1983.
[49] Taxas Instruments Incorporated,TMS320 User’s Guide,1983.
[50] R.J. Coy,C.N.Smith and P.R.Smith.HF-band radio receiver de-sign based on digitalsignal processing.Electronics & Communication Engineering Journal, April, 1992.
[51] D.T. Anderson , J.M . Whikehart. A digital signal processing HF re-ceiver.In:Proc. IEEE Conf. HF Communication Systems and Tech-niques,Pub.,February,1985,24.
[52]谈振藩.DSP上实现的卡尔曼滤波器[J].哈尔滨船舶工程学院学报,1987, 8(2):24-32.
[53]朱洪亮,葛临东.短波软件无线电接收机射频数字化的一种实现方法[J].单片机开发与应用.2006,(7):28-32.
[54]朱洪亮,葛临东,刘世刚.新型短波窄带接收机射频前端的设计与实现[J].信息工程大学学报,2006,5(2):16-25.
[55]汪烈军.一种基于双谱和复数谱方差信号检测算法研究.新疆大学硕士论文,2006.06.30.
[56]沈琰,薛继华,李跃进.一种中频数字化接收机的硬件实现[J].通信与广播电视,2005,3.
[57]杨彦昌.I COM-700TY电台数传接口改进.军事通信技术,1997,12,54-58.
[58]李双勋.TLV320AIC23在音频处理中的应用[J].国外电子元器件,2003,(10):45-49.
[59]程智辉.车载NVIS天线的研究.西安电子科技大学硕士论文,2008.10.01.
[2]高晋占.探空莫尔斯码自动识别[J].电子测量与仪器学报,2006.(8): 5-9.
[3]樊昌元,文斌.应用于民航空管的TPU莫尔斯译码器[J].电子测量与仪器学报,2004,(9):45-48.
[4]马威,张敬修,王虎帮.Morse电码自动译码系统[J].网络与信息技术,2007, 26(6):51-55.
[5]林瑞华.用微处理机对莫尔斯电码进行机器译码[J].通信与计算技术,1984,(3):126-128.
[6]高勤,都新学.莫尔斯信号的新型滤波方法[C].中国电子学会第七届学术年会论文集.2001.
[7]邓自立,王欣,高媛.建模与估计[M].北京:科学出版社,2007.
[8]皇甫堪,陈建文,楼生强.现代数字信号处理[M].北京:电子工业出版社,2003.
[9]易克初.语音信号处理[M].北京:国防工业出版社,2000.
[10]于宏毅,张贴等.手工莫尔斯报的一种新型自动收报算法及其手法识别算法[J].电子对抗,1995,11(4):317-320.
[11]岳喜才,郑崇勋.基于离散Gabor谱的短波电报信号检测.数据采集及信号处理,1995,11(4):317-320.
[12]林佳仕.短波通信中信号检测算法研究.哈尔滨工程大学硕士论文,2003.
[13]戴逸松.微弱信号检测方法及仪器.北京:国防工业出版社,1994,1.
[14]朱岩.强噪声下基于小波变换的语音增强研究.哈尔滨工程大学硕士论文,2003.
[15]王国权,薛申芳,金声震,宁书年.卫星自主导航中卡尔曼滤波算法改进与计算机仿真.计算机仿真,2004,(7):17-21.
[16]冯鹏,邹世开.卡尔曼滤波器UD分解滤波算法的改进.遥测遥控,2005.1.
[17]庄琼琼.多传感器组合卡尔曼滤波技术研究.北京航空航天大学硕士学位论文,2002.
[18]刘理.基于信号相位匹配原理的单频信号检测.西北工业大学学报,2005,(2): 35-41.
[19]孙轶源.利用信号相位匹配原理的正弦信号参数估计.自然科学发展,2001,11.
[20]杨祥龙,汪乐宇.一种强噪声背景下弱信号检测的非线性方法.电子与信息学报,1999,14(1):234-235.
[21]林佳仕.短波通信中信号检测算法的研究.哈尔滨工程大学硕士学位论文,2004.1.
[22]何立刚.噪声背景下短波莫尔斯信号的自动检测和识别研究.哈尔滨工程大学硕士论文,2005.
[23]袁文澹,唐伟.微机模拟自动拍发MORSE电报[J].计算技术与自动化,2001,2.
[24]耿振华.用单片机制制作多功能莫尔斯码电路[J].电子报,2005,05-08.
[25] [美]阿姆巴德.信号、系统与信号处理[M].冯博琴等译.机械工业出版社,2001.
[26]程佩青.数字信号处理教程[M].北京:清华大学出版社,2001.
[27]李跃进.短波电台销噪技术的实现方法[J].通信与广播电视,2006,(2):124-127.
[28]霍建,刘鸿雁,段秀铭.手工拍发的莫尔斯码信号种类识别算法设计[J].鞍山科技大学学报,2006,29(4):351-355.
[29]林金朝,李国军,周晓娜,曾理.强噪声背景下高频CW电报信号的自动检测[J].山西电子技术,2008,10.39-40.
[30]李春晓,赵旦峰,李强.用语音识别技术实现莫尔斯报的自动识别[J].信息技术.2006,(2):34-39.
[31]陈莹莹,彭仕勇,高勇.莫尔斯报务信号的检测和识别[J].计算机应用与软件,2009,26(6):241-243.
[32]潘德惠.卡尔曼滤波浅述[J].信息与控制,1982,(1):42-46.
[33]张正林,谭文生.从无线通信的发展历程看战术短波通信的技术体制和发展趋势[J].陕西烽火通信,2002,(1):1-2.
[34]张莉,张骏凌,龚耀寰.软件无线电数字中频技术[J].电讯技术,1999,(6):48-51.
[35]杨路刚,陈斌,王永斌,喻莉.基于MSP430的莫尔斯报训练装置设计[J].现代电子技术,2009,(9):8-10.
[36]高晋占.探空莫尔斯码自动识别[J].电子测量与仪器学报,2006,20(4):5-9.
[37]冯锡生.短波通信电路设计.北京:国防工业出版社,1985.
[38]王敬东,申敏.TI TMS320C55x DSP在TD-SCDMA系统中的应用[J].重庆邮电学院学报,2002,14(2):22-25.
[39]张汝波,何立刚,李雪耀.强噪声背景下莫尔斯信号的自动检测与识别[J].哈尔滨工程大学学报,2006, 27 (1): 113-117.
[40]王金龙,沈良,任国春等.无线通信系统的DSP实现[M].北京:人民邮电出版社,2002.
[41]周学科,张太镒,王小军.基于软件无线电的短波中频数字化[J].重庆邮电学院学报,2004,16(4):34-37.
[42]李双田,李昌立,陈丹平,袁津润.短波数字接收机的算法研究和DSP实现[J].电子学报,1999,(7):27-34.
[43] R. M. Lober.A DSP-based approach to HF receiver design:Higher Performance at a Lower Cost. TECHNICAL SYMPOSIUM,1993
[44] R.J.Coy, C.N.Smith and P.R.Smith.Use of DSP within a high performance HF band receiver.In:Proc IEE Conference of Radio Receivers and Associated Systems, July,1990.
[45] Qian Shie, Chen Dapang. Optimal biorthogonal analy-sis window function for discrete Gabor transform. IEEE Trans Signal Processing, 1994,4.
[46] MITOLA J. The software radio architecture [J].IEEE Communication Magazine,1995, 33.
[47] Castella,F.R.,An adaptive two-dimensional Kalman tracking filter,IEEE Trans. Aerosp, Election Syst.,Vol,AES-16,
[48] Taxas Instruments Incorporated,TMS320 Assembly Language Programmer’s Guide, 1983.
[49] Taxas Instruments Incorporated,TMS320 User’s Guide,1983.
[50] R.J. Coy,C.N.Smith and P.R.Smith.HF-band radio receiver de-sign based on digitalsignal processing.Electronics & Communication Engineering Journal, April, 1992.
[51] D.T. Anderson , J.M . Whikehart. A digital signal processing HF re-ceiver.In:Proc. IEEE Conf. HF Communication Systems and Tech-niques,Pub.,February,1985,24.
[52]谈振藩.DSP上实现的卡尔曼滤波器[J].哈尔滨船舶工程学院学报,1987, 8(2):24-32.
[53]朱洪亮,葛临东.短波软件无线电接收机射频数字化的一种实现方法[J].单片机开发与应用.2006,(7):28-32.
[54]朱洪亮,葛临东,刘世刚.新型短波窄带接收机射频前端的设计与实现[J].信息工程大学学报,2006,5(2):16-25.
[55]汪烈军.一种基于双谱和复数谱方差信号检测算法研究.新疆大学硕士论文,2006.06.30.
[56]沈琰,薛继华,李跃进.一种中频数字化接收机的硬件实现[J].通信与广播电视,2005,3.
[57]杨彦昌.I COM-700TY电台数传接口改进.军事通信技术,1997,12,54-58.
[58]李双勋.TLV320AIC23在音频处理中的应用[J].国外电子元器件,2003,(10):45-49.
[59]程智辉.车载NVIS天线的研究.西安电子科技大学硕士论文,2008.10.01.